[Talk-GB] worrying edit to junction (now reverted)

[Ken Kilfedder]

> I do think you are overreacting a little, the first edit was made by a
> new mapper, with 20 edits to her name now, this was her 18th. Don't we
> kind of expect new mappers to make mistakes? I still make mistakes,
> it's best to assume good intentions.

For an 18th edit, I thought it looked pretty good!  I won't have dared 
remodel a junction for min 18th edit.  The reason I assumed something was 
'up' was because of the similarly-named user who cam in and tidied up the
route relations.   I assumed the first user was a 'bot' and the second
was a bot herder.
 
> If this wasn't a Facebook mapper I doubt you'd have posted to the list,

Fair.


> To me, it looks like the second editor is some kind of internal
> validator, checking other Facebook mappers aren't breaking things. He
> caught some obvious mistakes, but didn't catch that the imagery is out
> of date.

This seems the most likely explanation.


> Maybe it'd be more productive to try and get them to change there
> process, to check osm history before making edits in highly mapped
> areas and objects that have a recent timestamp. (cc osm at fb.com)

Agreed; in fact the 'supervisor' guy has been quite chatty, and has
at least acknowledged my suggestion to look at the history for well-mapped
objects that don't match the imagery.  How well that filters through to
the rest of FB's teams we may never know.